In clinical samples the free, or non-protein bound, fraction of a drug or hormone is of great interest since it represents the compound's biologically-active form. This makes the measurement of this fraction ideal for patient diagnosis and treatment. But current analytical methods for these measurements (typically based on immunoassays) are slow, labor- intensive, and cannot be performed by existing automated systems. This is due to the low concentrations which must be determined and the need to separate the bound and free fractions of the analyte prior to analysis. In addition, the effects of bound analyte dissociation during the separation process are not well understood. The aim of this study is to obtain a better understanding of this dissociation and to develop new, automated systems for free drug/hormone measurements. Dissociation effects will be studied by examining the protein binding of two model compounds, thyroxine and cortisol. Systems for automating free drug/hormone assays for these compounds will be developed based on high-performance immunoaffinity chromatography and chemiluminescent detection (HPIAC/CL). In this method, an immobilized antibody column will be used to rapidly extract free analyte from a sample, followed by analyte detection using chemiluminescent acridinium ester labels. The first section of this work will examine the interaction kinetics of thyroxine and cortisol with their binding proteins. This will be done by studying the retention and band-broadening of their binding proteins as they pass through columns containing immobilized thyroxine or cortisol. The goal is to obtain a better understanding of dissociation effects in free drug/hormone assays and to provide data for minimizing these effects in HPIAC/CL systems. The second section will study the processes involved in the adsorption of analytes to immobilized antibody columns. The goal is to determine what optimum chromatographic conditions are needed for automating free drug/hormone immunoassays using HPIAC/CL. This will be studied using computer simulations and work with a model system, the binding of concanavalin A to fluorescent and nonfluorescent sugars. The last section will use data from the previous studies to design systems based on HPIAC/CL for free thyroxine and cortisol measurements. A comparison of this approach to current methods will be made to determine its potential advantages and limitations. The use of this system in measuring other drugs/hormones will then also be considered.